Bank Lending in Times of Large Bank Reserves∗ Antoine Martin,a James McAndrews, and David Skeieb aFederal Reserve Bank of New York bMays Business School, Texas A&M University Reserves held by the U.S. banking system rose from under $50 billion before 2008 to $2.8 trillion by 2014. Some econo- mists argue that such a large quantity of reserves could lead to overly expansive bank lending as the economy recovers, regardless of the Federal Reserve’s interest rate policy. In con- trast, we show that the amount of bank reserves has no effect on bank lending in a frictionless model of the current bank- ing system, in which interest is paid on reserves and there are no binding reserve requirements. Moreover, we find that with balance sheet costs, large reserve balances may instead be contractionary. JEL Codes: G21, E42, E43, E51. ∗This paper is a revision of Federal Reserve Bank of New York Staff Report No. 497, May 2011 (revised June 2013), and was previously circulated with the title “A Note on Bank Lending in Times of Large Bank Reserves.” The authors thank Todd Keister for valuable conversations that contributed to this paper; John Cochrane, Doug Diamond, Huberto Ennis, Marvin Goodfriend, Ellis Tallman, Steve Williamson, Steve Wolman, and the discussants of the paper, Morten Bech, Jagjit Chadha, Lou Crandall, Oreste Tristani, and Larry Wall, for helpful sugges- tions and conversations; the editor John Williams and two anonymous referees for very helpful comments; participants at the 2011 IBEFA/ASSA meetings, the 2013 Federal Reserve System Committee Meeting on Financial Structure and Regula- tion, the ECB conference on “The Post-Crisis Design of the Operational Frame- work for the Implementation of Monetary Policy,” the Swiss National Bank con- ference on Policy Challenges and Developments in Monetary Economics, and at seminars at the Bundesbank, the Federal Reserve Bank of New York, the Federal Reserve Board, and the Swedish Riksbank for helpful comments; and Sha Lu and particularly Ali Palida for outstanding research assistance. The views expressed in this paper are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors are the responsibility of the authors. Author e-mails: [email protected]; [email protected]; [email protected]. 193 194 International Journal of Central Banking December 2016 1. Introduction The amount of reserves held by the U.S. banking system rose from under $50 billion before 2008 to $2.8 trillion by 2014. This signif- icant increase is important because some economists and financial market participants claim that large levels of bank reserves will lead to overly expansive bank lending.1 Despite such concerns, little for- mal analysis has been conducted to show such an effect under the current banking system. In contrast, other commentators on the economy claim that the large level of reserves held in the banking system is evidence of a lack of bank lending. In this paper, we present a basic model of the current U.S. bank- ing system, in which interest is paid on bank reserves and there are no binding reserve requirements. We find that, absent any fric- tions, lending is unaffected by the amount of reserves in the bank- ing system. The key determinant of bank lending is the difference between the return on loans and the opportunity cost of making a loan. We show that this difference does not depend on the quan- tity of reserves. Moreover, when we introduce frictions, in the form of a cost related to the size of a bank’s balance sheet, increases in reserves may actually reduce bank lending and lead to a decrease in prices. The current banking system in the United States and world- wide no longer resembles the traditional textbook model of frac- tional reserve banking. Historically, the quantity of reserves supplied by a central bank determines the amount of bank loans. Through the “money multiplier,” banks expand loans to equal the amount of reserves divided by the reserve requirement. However, in many countries, reserve requirements have been reduced either to zero or to such small levels that they are no longer binding.2 1In speeches, former Federal Reserve Bank of Philadelphia President Charles Plosser expressed concern about the eventual need “to restrain the huge volume of excess reserves from flowing out of the banking system” (Plosser 2011), and former Federal Reserve Bank of Dallas President Richard Fisher cautioned about “excess reserves waiting to be converted to bank loans” (Fisher 2009). Meltzer (2010) expresses similar concerns. 2Bennett and Peristiani (2002) show that reserve requirements have been largely avoided in the United States since the 1980s by sweep accounts, and that the remaining reserve requirements are largely met by vault cash that banks hold Vol. 12 No. 4 Bank Lending in Times of Large Bank Reserves 195 Starting in the late 1980s, the Federal Reserve supplied the quan- tity of reserves needed to maintain its policy target—the federal funds rate—which is the interest rate at which banks lend reserves to each other in the interbank market. The Federal Reserve did not target the amount of reserves, the quantity of deposits or loans on banks’ balance sheets, or broad measures of the money supply. In that regime, the federal funds rate represents a bank’s alternative return on assets and hence is the required marginal return on bank lending. Banks expand their balance sheets so long as the marginal cost of funding is less than the marginal return on bank lending, abstracting from credit and liquidity risk. The federal funds rate sets the level of the required marginal return. From 2007 through 2014, the Federal Reserve greatly expanded the scope of its tools to address the financial crisis and a severe recession. Bank reserves increased rapidly during the financial crisis as the Federal Reserve provided unprecedented unsterilized lending through several facilities after the bankruptcy of Lehman Brothers. Reserves increased much further during the weak economic recov- ery as the Federal Reserve purchased Treasury securities, agency mortgage-backed securities, and agency debt as part of the large- scale asset purchases (LSAPs), also known as “quantitative easing,” or “QE.” Altogether, between September 2008 and mid-2014, bank reserves grew from under $50 billion to $2.8 trillion, as illustrated in figure 1. To allow the Federal Reserve to continue targeting its pol- icy rate even with large reserves outstanding, Congress accelerated previously granted authority for the Federal Reserve to pay interest on reserves in the Emergency Economic Stabilization Act of 2008. The Federal Reserve began paying interest on reserves on October 9, 2008. Paying interest on reserves allows the Federal Reserve to choose the required return on banks’ reserves independent of the quantity of reserves in the banking system.3 at branches and automated teller machines. As of mid-2008, required reserves were $71 billion, just 0.6 percent of total bank assets, and vault cash satisfied $43 billion of these requirements. Carpenter and Demiralp (2012) show empirically that the money multiplier does not hold using data from 1990–2007. 3For details and complementary analysis of interest on reserves as a monetary policy tool, see Ennis (2014), Ennis and Keister (2008), Keister, Martin, and McAndrews (2008), and Keister and McAndrews (2009). For details and analysis of additional new Federal Reserve monetary policy tools, including overnight 196 International Journal of Central Banking December 2016 Figure 1. Large Quantity of Reserves in the Banking System Source: Federal Reserve statistical release H.4.1: Factors Affecting Reserve Bal- ances. Notes: Frequency: biweekly. Reserve balances with Federal Reserve Banks are the difference between “total factors supplying reserve funds” and “total factors, other than reserve balances, absorbing reserve funds.” This item includes balances at the Federal Reserve of all depository institutions that are used to satisfy reserve requirements and balances held in excess of balance requirements. It excludes reserves held in the form of cash in bank vaults, and excludes service-related deposits. We introduce a new framework in which the role of fiat reserves that pay interest can be studied in a general equilibrium bank- ing economy with a closed system of bank payments and central bank reserves. We include banking, corporate, and retail sectors, which transact in competitive markets for bonds, deposits, loans, and goods. Our benchmark model shows that, without frictions, bank lending quantities and interest rates are invariant to the level of reserves chosen by the central bank. Banks lend up to the point where the marginal return on lending equals the return on holding reverse repurchases (reverse “repos”) and the term deposit facility (TDF), see Martin et al. (2013), which expands upon the modeling framework in this paper. Vol. 12 No. 4 Bank Lending in Times of Large Bank Reserves 197 reserves, which is equal to the interest rate on reserves set by the central bank. This provides an indifference result for the quantity of reserves. In particular, while the size of banks’ balance sheets expands with increases in reserves, all else equal, the lending deci- sion for a bank is determined by the same marginal return condition as with the former method of monetary policy implementation. A loan is made at the margin if its return exceeds the marginal oppor- tunity cost of reserves, whether that is the federal funds rate as in the prior regime or the rate of interest on reserves as in the current regime. We also demonstrate that the quantity of reserves held in the banking system in the absence of binding reserve requirements or significant currency withdrawals is determined in the United States solely by the Federal Reserve.